malloc: Trim unused arenas on thread exit

[Florian Weimer]

On 11/13/2017 10:57 PM, DJ Delorie wrote:
> Florian Weimer <fweimer@redhat.com> writes:
>> I think trimming unused arenas as much as possible is a good heuristics
>> to minimize RSS, so getting this to work might be worthwhile.
> 
> I ran this against many of my real-world workloads and did not see a
> significant (i.e. obvious) reduction in actual RSS.  Some tests showed a
> minor (i.e. untrustworthy, I didn't run enough passes to reduce the
> noise below any measured change) *increase* in RSS, as well as an
> expected (but still untrustworthy) increase in cycles required.

Sorry, as I tried to say in my original message: It does not actually 
reuse RSS because the sub-heaps are not deallocated.  I don't know why 
this does not happen.

The trimming really should be a win because the workload is likely to 
change when a thread exits.

Thanks,
Florian